By Shammy Narayanan
It’s easier to spot a teenager without a Tinder account than an organization without a centralized Data team. Unfortunately, the hype around Data is so overrated that companies scurry to create the Data teams just as Ostriches hurry to put their head in the sand. Such ill-formed teams are like the cart before a horse; they effectively function in silos but struggle to permeate and coordinate with the rest of the product areas. Hence they often get swiped left and teeter on the verge of break-ups before a management-mediated temporal patch-up. Forming a centralized Data team an overarching idea or an outdated one?
Monolithic Architecture: Monoliths are the meths of Application Architecture; none seems to like them, yet we aren’t successful in eliminating them. When we have not even at the halfway mark in crossing this Chinese wall of remediating the code Monoliths, why are we rapidly creating a Data Monolith by assembling a centralized Data team? are we not learning from our past? to quote Einstein, “Insanity is doing the same thing over and over and expecting different results.”
Business Innovation: A set of thousand line items in a relational database will be viewed as a thousand members/accounts by a Product area, whereas the Data team will technically view it as one thousand records that need to be curated, transformed, and ready for the subsequent phase. This difference in view stems from the fact that the Data team is accountable for managing millions of records, thousands of Tables, Hundreds of Databases, and a bunch of Dashboards. With such a tremendous load on their shoulders, expecting them to view each transaction purely from a domain perspective is akin to requesting a butcher to exercise sympathy toward the sheep; the job will never get done. However, business values/innovation cannot be sacrificed at the altar to gain brute efficiency. This balance can be accomplished by having data engineers and analysts as part of the product teams rather than as a centralized one. Evolving “Data Mesh” is a strong proponent of de-centralized groups enabling organic collaboration and fostering Business innovation.
Profit Models: Even Financially, it makes more sense to embed Data into the Product teams. Traditionally Data Teams are often treated as Cost Centres while Product teams are as Profit centers. When we nurture an aggressive ambition to leverage data as differentiators and identify possible new revenue opportunities, it’s ironic to continue Data as part of cost centers that are highly vulnerable to cost-cutting and first in line to get hit by Industry slowness. It’s akin to cutting the limb and taking up a driving job !!
This dilemma about “Centralized or Federated” Data teams doesn’t have a cookie-cutter response; it’s a function of organizational maturity. A centralized model is a foundational step; this will help to identify, establish and refine the scope, process, guidelines, and, more essentially, harvesting niche data talents. When the journey commences here, it shouldn’t end but evolve. The Federated model is the next, the Product teams have an embedded data component similar to the Agile team having a functional tester (as against the centralized functional testing team in the waterfall model). Certain non-negotiables, such as Data Privacy ( e.g., GDPR) security, Data Governance, and Cross product features, will require a representative(s) from product teams to come together to establish and implement enterprise guidelines. It’s absolutely fine if you are not in an entirely federated model, but as long as you have started marching, it’s progress.
Data as a Product is pretty new and are still an evolving concept with many uncertainties and unknowns. When we can learn a lot by looking at the past; however, it’s not a great idea to drive on a freeway by focusing only on the rear-view mirror; nevertheless, a few risks had to be taken, lessons to be learned, and amends to be made; such changes are an inherent part of any promising journey, once this foundation is laid, your Data Team will be ready to be deliver the Killer Punch.
Narayanan is the Data and Analytics Head of a Healthcare GCC. He is 9x certified in Cloud and blogs about emerging tech and strategies. He can be reached at https://www.linkedin.com/in/shammy-narayanan-9a098518/